Google vs. Amazon: A Choice, Not An Echo

Amazon, the online retailer, is emerging as Google’s most effective competitor in the world of web-based applications – and NOT by going head-to-head with the search monster, at least, not now.

No one is likely to beat Google playing Google’s game. Which is why Amazon, rather than MSN or Yahoo, is so interesting. Instead of Google’s advertising supported free-for-all approach, Amazon generally brings a very different model to the web:

Each company’s respective storage services offer a clear example of the differences. Note that since the rumored Gdrive storage service is still a rumor, I’m comparing Gmail’s storage capability to Amazon’s S3 (Simple Storage Service) service, where the differences are illustrative, not definitive. For those interested Gmail sync and Gmail drive shell storage hacks are available.

S3 has some interesting design requirements and principles a few of which are:

• 99.99% Availability – enterprise snobs may sniff but less than an hour’s downtime per year is quite good
• Scales in capacity, performance, I/O rate and users to support an unlimited number of applications
• Commodity Hardware using software that expects and handles frequent hardware failures with no or minimal interruption
• Limited concurrency to minimize needed controls and ACID-style monitors
• Service Decomposition into small, understandable building blocks that may be re-used for other services

There is more to what Amazon is doing than S3. I’ll be covering a number of cool topics about Amazon, Service-Oriented Architectures (SOA) and lightweight web applications in future posts.

Two threads StorageMojo.com will be covering with particular care over the next few weeks.

• Amazon’s architecture
• The impact of SOA and lightweight web apps on infrastructure. Hint: think huge.